McCauley was determined to conduct an exhaustive investigation. As a former prosecutor, she knew the bar would be high for the Suffolk County district attorney’s office to prove criminal behavior beyond a reasonable doubt. She also knew she had more police work to do. But based only on what she and her partner had already uncovered, she had no doubt about their ability to demonstrate probable cause, which the DA’s office would require in order to file criminal charges. As far as which players should be held to account, Harald Grob remained first on her list. She continued to believe, however, that the list should probably be longer than just one name.
McCauley hadn’t spoken to the Norwesco project manager since his interview two days after the accident, and he had returned to Canada shortly after that. In the five months since, however, McCauley had uncovered mounds of fresh material as well as several shocking findings that made her question key aspects of Harald’s account.
Still, McCauley worried how a potential prosecution against Harald might play out in court. After all, during his interview he had stressed that all the major parties—the MWRA, Kiewit, Kaiser, and PB—had not just approved his plan but had weighed in with suggestions for refining it. Given the involvement of all those big players, it would be hard to paint Harald as some kind of rogue actor. With all that shared responsibility and liability, then, the whole diving operation resembled the way patients are often cared for at the highly respected but labyrinthine teaching hospitals that made Boston a medical mecca. The patients see so many interns and residents and fellows and attending doctors and department chiefs and specialists, all swooping in and out of their room, that it’s impossible to figure out who’s in charge of their care. Everyone’s in charge, so no one’s in charge.
On top of all that, McCauley worried about the prospect of Harald taking the stand and pointing his finger at OSHA for having failed to raise any objections about his plan, when the agency had received it well ahead of time. She asked herself, will that play to a jury? She knew the answer. Not well at all.
Two weeks into the new millennium, McCauley finished her twenty-four-page report on the sudden deaths of Billy and Tim. The nearly six-month investigation she had conducted with the help of her state police colleagues288 had sometimes felt to her like a series of intense graduate school seminars. The surviving divers had been among her most important tutors, impressing her with their recall and their candor, even when it came to agonizing details, such as what had happened to Tim’s foot. The detective’s other tutors included equipment suppliers, project managers, and especially the expert analysis of the navy’s Dr. Knafelc. In writing her report, McCauley leaned heavily on the navy analysis as she worked to answer all the big questions that had confounded her when she’d started.
How responsible for the disaster was the seemingly slapped-together liquid gas breathing system? Knafelc had answered that question head-on in her report. “The design of the breathing apparatus289 was inadequate in its ability to support working divers,” her analysis read. “Lack of appropriate monitoring equipment contributed to the death of two individuals.”
The navy report faulted Harald for giving the divers the Multiwarn handheld gas analyzers rather than installing a far more reliable in-line analyzer, as Riggs had suggested. For short money290, that plumbed-in analyzer would have provided rock-solid, continuous monitoring of the breathing supply and would have given Billy and Tim fair warning of impending doom. Knafelc noted Harald’s failure291 to account for the predictable loss in air pressure that took place at each point where the five different three-hundred-foot umbilicals were connected. And the navy doctor complained that Harald had failed to test the breathing system adequately, under full-load conditions—again, something Riggs had suggested—before sending the divers into the tunnel. Those light walk-through simulations that Harald had instructed the divers to do in Tap’s yard had been a poor substitute, especially given the absence of the key component of the system, the MAP Mix 9000.
What explained the Sno-Cone effect, the frost that enveloped the liquid oxygen tank and regulator? As the divers worked harder at the end of the tunnel, that put a greater demand on the mixed-gas system back at the Humvee. That increased demand, in turn, required the liquid tanks to produce more gas—to vaporize more product. Liquid gas is extremely cold—cold enough, of course, to burn warts off skin—because the cold allows the molecules to slow down so they can be packed more tightly into the tank. But in the process of being vaporized292 into gas, those molecules expand rapidly. To do that, they pull heat in from the surrounding environment as their source of energy. As cold as the tunnel was, it was a whole lot warmer than inside those cylinders of liquid gas.
As the liquid oxygen cylinder struggled to keep up with the divers’ heightened demand, those molecules raced to expand more quickly. Because they needed more energy to do that, they continued to draw in whatever heat they could get from the outside environment, beginning with the regulator sitting atop the cylinder. That made the regulator colder and colder. And because the tunnel environment was so damp, as the regulator got colder, the moisture on it began to turn to frost. To keep expanding, those molecules of gas inside the cylinder intensified their demand for more warmth from the outside. Eventually that turned the frost covering the regulator into ice. That process kept going, layering more and more ice onto the regulator, until it took on the appearance of an “ice sculpture,” as Riggs had so memorably put it. By then, it was about as cold on the surface of the cylinder as it was inside it. The regulator was as frozen as an outdoor faucet in a blizzard, unable to allow anything to get by. It simply gave out. With no more oxygen coming through, the divers’ mixed-gas breathing air basically turned into a pure shot of nitrogen.
Judging from Harald’s lengthy resume, the navy doctor was stunned that he wouldn’t have known the added risks involved with this kind of system. “Using liquid gases requires293 closer attention to detail because of the potential for the regulator to freeze,” the navy analysis read. “The designer’s experience in diving, including nitrox and mixed-gas, and training as a professional engineer should have provided him the necessary tools to design and build a safe, effective breathing gas system.”
The navy doctor also faulted Harald for failing to include two fairly basic but essential components, which could have gone a long way toward guarding against the Sno-Cone effect. The first missing device was a volume tank, essentially a way station where the breathing gas could have been stored after the vapors of the liquid nitrogen and liquid oxygen had been blended. That volume tank would have allowed the mixed gas to be tested before it ever reached the divers’ facemasks. It would also have buffeted the cylinders of liquid gas from the varying demands of the divers, reducing the chance of the regulator freezing over. Breathing directly off a mixer294 “with only episodic gas sampling is not acceptable,” the navy report noted.
The second missing device295 was a heat exchanger. This component would have helped the system safely deliver the breathing air at a higher flow rate as the divers’ demand increased. It was mystifying why the divers’ system had no heat exchanger even though the Humvees did. That tall metal device, the one that resembled the radiator in a trendy condo, had ensured that the liquid oxygen tank on the back of the Humvee was converted into gas at the temperature needed to keep the vehicle humming. The divers deserved at least as much.
288 (Chapter 14, note 9)
with the help of her state police colleagues: McCauley stressed the key roles played by her partner, Trooper John O’Leary; their supervisor, Sgt. Randy Cipoletta; and Trooper Wes Wanagal of the State Police Crime Scene Services Section.
289 (Chapter 14, note 10)
“The design of the breathing apparatus”: Dr. Marie E. Knafelc, “Investigation of Deaths Within the Boston Harbor Outflow Tunnel,” Navy Experimental Diving Unit, September 3, 1999, 6.
291 (Chapter 14, note 12)
Knafelc noted Harald’s failure: Ibid., 3-4. The twelve-hundred-foot umbilical connecting the manifold in the Humvee to the manifold in the boat consisted of four three-hundred-foot hoses strung together. From there, each diver on the excursion team had his own three-hundred-foot umbilical between him and the manifold in the boat.
292 (Chapter 14, note 13)
But in the process of being vaporized: Dr. Marie E. Knafelc, interviews by author, May 2012; Canadian Centre for Occupational Health and Safety, “Cryogenic Liquids—Hazards,” www.ccohs.ca/oshanswers/chemicals/cryogenic/cryogen1.html; and Air Products, “Cryogenic Liquid Containers,” Safetygram-27, www.airproducts.com/~/media/Files/PDF/company/safetygram-27.pdf.
293 (Chapter 14, note 14)
“Using liquid gases requires”: Knafelc, “Investigation of Deaths,” 5.